INDEX

Explanations

instances of Spanish text with varying activations, especially related to organizational structures, political figures, and specific locations

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-ajt

Prompts (Dashboard)

12,288 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

46,080

Data Type

torch.float32

Hook Point

blocks.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

apollo-research/Skylion007-openwebtext-tokenizer-gpt2

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

•Layer 6 UMAP region: Mostly-local line at bottom - local

No Comments

Negative Logits

velt

-0.50

 stride

-0.50

pedia

-0.50

 tide

-0.50

 tires

-0.50

 signalling

-0.49

Shell

-0.48

kered

-0.48

 tire

-0.48

 locom

-0.47

POSITIVE LOGITS

BIL

0.59

gage

0.57

Ĭ±

0.55

ellar

0.53

gmail

0.53

phia

0.53

ruction

0.52

ogyn

0.51

atl

0.51

aternity

0.50

Activations Density 17.332%

instances of Spanish text with varying activations, especially related to organizational structures, political figures, and specific locations

No Comments

No Known Activations